The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external goals and brand facts to generate social media content, creating a surface for indirect instructions in SKILL.md.
Ingestion points: Data enters the context via user goals in jobs_create and brand positioning data in brain_write (SKILL.md).
Boundary markers: The skill includes specific ground rules requiring the agent to review all drafted content and obtain verbal user confirmation before calling 'live' tools like publishing or approving (SKILL.md).
Capability inventory: The agent has capabilities to draft, schedule, and publish content to linked social platforms, and modify persistent brand memory (SKILL.md).
Sanitization: The platform API performs validation and returns structured refusal codes (e.g., 422) when content violates platform-specific rules (SKILL.md).
[EXTERNAL_DOWNLOADS]: The skill references documentation and executes a package runner from the service provider's infrastructure.
Evidence: The skill fetches Markdown documentation and OpenAPI specifications from shapelessai.com and utilizes the shapelessai package via npx.

shapeless